regional versus multilateral trade ... - economics
TRANSCRIPT
Regional versus Multilateral Trade Liberalization,
Environmental Taxation and Welfare
Soham Baksi
Department of Economics Working Paper Number: 2011-03
THE UNIVERSITY OF WINNIPEG
Department of Economics
515 Portage Avenue
Winnipeg, R3B 2E9
Canada
This working paper is available for download from:
http://ideas.repec.org/s/win/winwop.html
Regional versus Multilateral Trade Liberalization,Environmental Taxation and Welfare
Soham Baksi �
AbstractWe consider strategic trade among identical countries and compare the impacts of
multilateral versus regional tari¤ reduction on equilibrium pollution tax and social
welfare. While both forms of trade liberalization increase production and consumption
in the tari¤-reducing countries, regional trade liberalization also reduces production
in a non-participating country and may decrease its consumption. When pollution is
local, regional and multilateral trade liberalization have similar impacts in the tari¤-
reducing countries. In contrast, when pollution is perfectly transboundary, regional
(multilateral) trade liberalization (i) weakens (may strengthen) environmental protec-
tion in the tari¤-reducing countries, and (ii) in the neighbourhood of free trade, may
increase (decreases) welfare of the tari¤-reducing countries.
JEL classi�cations: Q58; F18; H23
Keywords: Environmental taxation; Trade liberalization; Regional trade agreements
�Department of Economics, University of Winnipeg, 515 Portage Avenue, Winnipeg, Canada R3B
2E9. Phone: 1 204 2582945. Fax: 1 204 7724183. Email: [email protected]
1
1 Introduction
Environmentalists have long worried that expanded international trade will detrimen-
tally and sometimes irreparably a¤ect the ecosystem. One concern is that competitive
pressures unleashed by trade will force governments to weaken their environmental
policies, and may even lead to the emergence of pollution havens. The transboundary
nature of many pollutants will also make it unlikely that governments, acting non-
cooperatively, will pursue globally e¢ cient environmental policies. A growing body of
literature in economics has examined these concerns (e.g. Krutilla, 1991; Ulph, 1996;
Antweiler, et al., 2001; Greenstone, 2002; McAusland, 2008; Managi, et al. 2009).1 In
general, economists have pointed out that the relation between international trade and
the environment is a multifaceted one and the impacts can go either way depending
on the speci�c context. For example, while the �scale e¤ect� associated with trade
liberalization tends to increase pollution, the �technique e¤ect� tends to decrease it.
Moreover, the �composition e¤ect�of trade liberalization can increase or decrease pol-
lution in a country depending on its comparative advantage (Copeland and Taylor,
2003).
In the presence of imperfect competition, countries trading with each other have an
incentive to strategically distort their environmental policy in order to increase their
welfare in the non-cooperative equilibrium (Barrett, 1994; Kennedy, 1994). When
imperfectly competitive �rms compete in terms of quantities, reducing domestic �rms�
environmental costs makes them more competitive internationally, enabling a country
to capture additional rents. This tends to ine¢ ciently weaken environmental policy in
these countries. At the same time, a desire to shift polluting production from itself to
its trading partners can lead each country to make its environmental policy ine¢ ciently
stringent. For the case of symmetric countries and transboundary pollution, Kennedy
(1994) concludes that the net impact under free trade is a lowering of the pollution tax
below its e¢ cient level.
While Kennedy identi�es how the equilibrium pollution tax may be distorted under
free trade, Burguet and Sempere (2003) examine how trade liberalization (in the form of
bilateral tari¤ reduction) a¤ects environmental policy and welfare by changing the var-
ious distortionary forces. Using a model of bilateral trade with imperfect competition
1Most of the literature on trade and environment has modeled pollution as a¤ecting the utilityfunction rather than the production function. A few papers that analyze the latter scenario includeCopeland and Taylor (1999), and Benarroch and Thille (2001).
2
and local pollution, the authors show that trade liberalization can make environmental
policy more or less stringent, depending on factors such as the convexity of the damage
function and the emission intensity of output. On the one hand, by increasing output,
trade liberalization increases marginal social cost of output, which tends to tighten
environmental policy. On the other hand, lower tari¤s imply lower import revenue,
which tends to make environmental policy more lax. The net impact on equilibrium
environmental policy depends on the relative strength of these counteracting forces.
Further, Burguet and Sempere �nd that when the environmental policy instrument
is a pollution tax, marginal social cost is always less than price. Consequently, by
increasing output, a bilateral tari¤ reduction always increases welfare of each country.2
One limitation of the literature on trade liberalization and strategic environmental
policy is that it usually considers a two-country framework and examines the symmet-
ric case of trade liberalization by both countries. This is akin to an examination of the
impact of multilateral trade liberalization (when there are more than two countries) on
the environment. However, much trade liberalization worldwide has involved regional
trade agreements, where a sub-group of trading countries decides to gradually reduce
tari¤ on each others imports.3 In fact, while multilateral trade negotiations under the
auspices of the WTO have fumbled in recent years, regional trade agreements (RTAs)
have proliferated.4 When tari¤ reduction takes such a regional form, the impact on
strategic environmental policy and welfare in the participating and non-participating
countries is likely to be di¤erent from those under multilateral tari¤ reduction. The
current paper examines this under-explored issue of the environmental and welfare
impacts of RTAs and compares these impacts with those under multilateral trade lib-
eralization.
While there has been a recent move towards incorporating environmental issues in
international trade agreements, the coordination of trade and environmental policies
2In their Proposition (p. 31), Burguet and Sempere (2003) note, �If the environmental instrumentis a tax (either on output, input, or emissions), a bilateral reduction in tari¤s increases welfare.�
3For example, under NAFTA, the participating countries (Canada, Mexico and the US) agreed toreduce their tari¤s in equal annual stages over a speci�ed number of years.
4As the WTO notes, �The surge in RTAs has continued unabated since the early1990s. As of 15 May 2011, some 489 RTAs, counting goods and services noti�ca-tions separately, have been noti�ed to the GATT/WTO. At that same date, 297 agree-ments were in force... The overall number of RTAs in force has been increasingly steadily,a trend likely to be strengthened by the many RTAs currently under negotiations.� (seehttp://www.wto.org/english/tratop_e/region_e/region_e.htm). For detailed information on all theRTAs currently in force, see http://rtais.wto.org/UI/PublicMaintainRTAHome.aspx
3
across countries remains largely absent (OECD, 2007).5 On the other hand, as interna-
tional trade agreements have progressively left countries with less �exibility to pursue
independent tari¤ policies, this has motivated some countries to use environmental
(and other domestic) policy instruments in order to achieve trade policy objectives
(Ederington and Minier, 2003).
When environmental policy is strategically chosen by countries, our analysis indi-
cates that the impact of freer trade on environmental protection and welfare depends
on the nature of the trade liberalization process. The di¤erences in outcomes for multi-
lateral vs. regional trade liberalization exist even when the countries are identical. We
�nd that, while multilateral trade liberalization increases production and consumption
in the tari¤-reducing countries, regional trade liberalization reduces production in the
non-participating country and may decrease its consumption as well. The changes in
consumption and production have corresponding impacts on the rent capture and pol-
lution shifting e¤ects in these countries. Further, the tari¤ e¤ect of trade liberalization
tends to reduce equilibrium pollution tax in tari¤-reducing countries in the multilateral
case, but tends to increase the non-participating country�s pollution tax in the case of
regional trade liberalization.
The paper shows that multilateral and regional trade liberalization a¤ects pollution
tax and welfare in tari¤-reducing countries in similar ways when pollution damage is
localized, but in dissimilar ways when pollution is largely transboundary. For global
pollutants such as greenhouse gases, we �nd that regional (multilateral) trade liber-
alization (i) weakens (may strengthen) environmental protection in the tari¤-reducing
countries, and (ii) in the neighbourhood of free trade, may increase (decreases) welfare
of the tari¤-reducing countries. The asymmetry inherent in regional trade liberalization
makes it more likely than multilateral trade liberalization to weaken environmental pro-
tection in the tari¤-reducing (but otherwise identical) countries. Furthermore, regional
trade liberalization is found to a¤ect pollution tax and welfare in the participating and
non-participating countries in dissimilar ways. For instance when pollutants have very
harmful but localized impacts, regional trade liberalization strengthens environmental
5The OECD (2007) report notes, �In spite of these developments, the number of RTAs includingsigni�cant environmental provisions remains small, and there is still much scepticism, especially amongdeveloping countries, toward dealing with environment in the context of trade agreements... Trade andenvironment debates have traditionally seen developing country negotiators cautious about incorporat-ing environmental considerations into multilateral trade agreements. Similar concerns apply to the in-tegration of environmental considerations into RTAs. Among the concerns are that environmental con-siderations will result in trade barriers or that their implementation will constitute an excessive burdenin terms of �nancial and human resources.�(http://www.oecd.org/dataoecd/56/55/38664937.pdf)
4
protection in participating countries but weakens such protection in non-participating
countries.
We proceed as follows. Section 2 presents the model, where three countries with
oligopolistic markets engage in international trade in a polluting good, and derives
its equilibrium. In section 3, we analyze the benchmark scenario where the countries
undertake multilateral tari¤ reduction. Section 4 then takes up the case of regional
trade liberalization where two countries (the �participating countries�) engage in bi-
lateral tari¤ reduction while the third country (the �non-participating country�) stays
out of this process. The results from multilateral and regional trade liberalization are
compared in section 5, while the last section concludes.
2 The model
Consider three identical countries, 1, 2 and 3, with segmented markets. There are n
�rms in each country, with n � 1. All �rms produce a homogeneous good and face aconstant marginal cost of production, given by c: Each �rm in country 1 sells xi units
of the good in country i; where i = 1; 2; 3. Each �rm in country 2 sells yi units in
country i. Each �rm in country 3 sells zi units of the good in country i.
In each market, �rms compete in quantities, i.e. à la Cournot. Inverse demand in
each country is identical and given by
p (Qi) = p (0)�Qi
where p (0)� c � a > 0; and Qi is total quantity sold in country i. Country 1 (respec-tively, country 2) charges a tari¤ of � per unit of import from country 2 (respectively,
country 1). Countries 1 and 2, each charges a tari¤ of � per unit of import from
country 3. Country 3 charges a tari¤ of � per unit of import from countries 1 and 2
(see Figure 1). The tari¤s are given exogenously in our model. Initially, i.e. pre-trade
liberalization, the tari¤ rates are the same (i.e. � = �) in identical countries.
Trade liberalization takes the form of a reduction in the relevant tari¤ rate. Two
alternative forms of trade liberalization are considered. We de�ne �multilateral trade
liberalization�as an equal multilateral reduction in the tari¤ rates � and � by all three
countries, and analyze this benchmark scenario in Section 3. The next section considers
�regional trade liberalization�, de�ned as an equal bilateral reduction in the tari¤ rate
� by countries 1 and 2 only (with � remaining unchanged). This re�ects the situation
5
subsequent to the signing of a regional trade agreement amongst a sub-group of the
trading partners (countries 1 and 2) while others remain outside the agreement (country
3, in our case). Unlike multilateral trade liberalization, regional trade liberalization
introduces an asymmetry among the three ex-ante identical countries.
A by-product of production in this industry is pollution. With appropriate de�ni-
tion of units, it is assumed that, for every unit of output produced, the �rms emit one
unit of pollution.6 The pollution is transboundary where 2 [0; 13] fraction of pollution
generated in one country a¤ects each of the other two countries. The social cost of
pollution is increasing and convex in the level of emissions a¤ecting a country. The
pollution damage function in country i; given by Di; is as follows:
D1 =1
2� [(1� 2 )X + Y + Z]2 (1)
D2 =1
2� [ X + (1� 2 )Y + Z]2 (2)
D3 =1
2� [ X + Y + (1� 2 )Z]2 (3)
where � � 0 is the pollution damage parameter, and the total production undertaken incountries 1, 2 and 3 are respectively X � n
P3i=1 xi; Y � n
P3i=1 yi and Z � n
P3i=1 zi:
Di¤erent values of the transboundary pollution or spillover parameter, ; allow us to
consider a continuum of cases ranging from strictly local pollution ( = 0) to perfectly
transboundary pollution ( = 13).7 The environmental policy in country i is a tax
imposed on domestic �rms at the rate ti per unit of emission. Given our assumption
of constant emission-output ratio, a tax per unit of emission is equivalent to a tax per
unit of the polluting good.
The sequence of moves is as follows. In the �rst stage, (an environmental authority
in) each country chooses its pollution tax to maximize the country�s own welfare, taking
the other countries�pollution taxes (and the tari¤s, � and �) as given. In the second
stage, each �rm takes the policies set by the countries and the output decisions of the
6Thus we are assuming a constant emission intensity of output. Emissions can be reduced througha reduction in output, with foregone pro�t being the abatement cost for a �rm. Allowing �rms toseparately choose their level of abatement does not change our results qualitatively.
7When = 13 , pollution generated in each country a¤ects all three countries equally. We are
ignoring situations where 2 ( 13 ;12 ], i.e. where pollution generated in each country a¤ects that
country less than it a¤ects the other countries. In a setting where pollution �ows from an upstreamto a downstream country, could exceed 1
3 in the upstream (but not the downstream) country. Sincewe are assuming is identical across the three countries, we restrict its value to the [0; 13 ] interval.
6
(3n� 1) other �rms as given, and chooses its own output. To obtain the subgameperfect Nash equilibrium, the model is solved using backward induction.
2.1 Second stage: Output decision of �rms
Total quantity sold in country i is given by Qi = n (xi + yi + zi). The pro�t maximiza-
tion problems of each �rm in countries 1, 2 and 3 are respectively given by (4) ; (5)
and (6):
maxx1; x2; x3
3Xi=1
xi (a�Qi)� t13Xi=1
xi � �x2 � �x3 (4)
maxy1; y2; y3
3Xi=1
yi (a�Qi)� t23Xi=1
yi � �y1 � �y3 (5)
maxz1; z2; z3
3Xi=1
zi (a�Qi)� t33Xi=1
zi � �z1 � �z2 (6)
The Cournot-Nash equilibrium quantities for the three markets are computed. The
quantities sold in country 1 and country 2 are:
x1 =1
3n+ 1[a+ n (�+ �)� t1 (1 + 2n) + n (t2 + t3)]
y1 =1
3n+ 1[a+ n�� (1 + 2n) (t2 + �) + n (t1 + t3)]
z1 = z2 =1
3n+ 1[a+ n�� (1 + 2n) (t3 + �) + n (t1 + t2)]
x2 =1
3n+ 1[a+ n�� (1 + 2n) (t1 + �) + n (t2 + t3)]
y2 =1
3n+ 1[a+ n (�+ �)� t2 (1 + 2n) + n (t1 + t3)] (7)
The quantities sold in country 3 are given by:
x3 =1
3n+ 1[a� � (1 + n)� t1 (1 + 2n) + n (t2 + t3)]
y3 =1
3n+ 1[a� � (1 + n) + n (t1 + t3)� t2 (1 + 2n)]
z3 =1
3n+ 1[a+ 2n�� t3 (1 + 2n) + n (t1 + t2)] (8)
7
We assume that parameter values are such that each of the above quantities is positive.
Country 1�s total production is X = n (x1 + x2 + x3), total consumption is Q1 =
n (x1 + y1 + z1), and net import is n (y1 + z1 � x2 � x3) = n�2t1 � t2 � t3 + n(���)
3n+1
�.
Similarly, country 3�s net import is n (x3 + y3 � z1 � z2) = n�2t3 � t1 � t2 + 2n(���)
3n+1
�.
2.2 First stage: Equilibrium environmental policy
In the �rst stage of the game, each country chooses the pollution tax rate that maxi-
mizes its own welfare, taking as given the tari¤ levels and the other countries�pollution
taxes. Social welfare is de�ned to be the sum of consumer surplus; producer surplus;
tari¤ revenue and pollution tax revenue less pollution damage:Welfare in country i,Wi;
is thus given by
Wi (t1; t2; t3) � CSi + PSi + TRi + ERi �Di; i = 1; 2; 3 (9)
In country 1, for example, we have that consumer surplus CS1 = 12(n (x1 + y1 + z1))
2 ;
producer surplus PS1 = n (x21 + x22 + x
23) ; tari¤revenue TR1 = n (�y1 + �z1), pollution
tax revenue ER1 = nt1 ( x1 + x2 + x3) ; and pollution damage D1 is as given by (1).
The non-cooperative Nash equilibrium pollution tax in each country is computed
by simultaneously solving the three �rst order conditions: @Wi(t1;t2;t3)@ti
= 0 for i = 1; 2; 3.
We denote these equilibrium taxes, which are a function of the tari¤rates, as t1 (�;�) =
t2 (�;�) and t3 (�;�). The second order condition for welfare maximization is satis�ed
since we have the following:
@2W1
@t21=@2W2
@t22=@2W3
@t23= �n212n+ 5 + 9� ( (6n+ 2)� 2n� 1)
2
(3n+ 1)2< 0.
3 Multilateral trade liberalization
We begin by analyzing the benchmark case of multilateral trade liberalization, where
all three countries start from the same initial tari¤ rate (i.e. � = �) and undertake
an equal marginal reduction in that rate.8 We analyze the impact of such multilateral
tari¤ reduction on the equilibrium pollution tax and welfare in each country. Since
8The benchmark case is similar to Baksi and Ray Chaudhuri (2009), who examine the e¤ects ofmultilateral trade liberalization using a two-country model and an alternative speci�cation of thepollution spillover process.
8
tari¤ rates are equal across the countries both before and after trade liberalization, the
equilibrium pollution tax rate and welfare levels are also identical in the three countries
under this symmetric scenario.
By substituting � = � in ti (�;�), the equilibrium pollution tax rate in each country
is derived as tmi = tm(�) for i = 1; 2; 3, where
tm(�) � 3a (2n+ 1)� 3�n (3a� 2�) (1� 2 + 2n (1� 3 ))� 2� (3n+ 3n2 + 1)
9n (2� (3n+ 1)� (1 + �) (2n+ 1))(10)
and the superscript �m�is used to denote this multilateral trade liberalization scenario.
The various sources of distortions that tend to make the equilibrium tax globally
ine¢ cient are as follows. First, there is the �transboundary externality e¤ect� that
tends to lower tm from its globally e¢ cient level, as each country ignores the impact of
pollution generated within its own borders on welfare in the other countries. Second,
there is the �rent capture e¤ect�that also tends to lower the equilibrium pollution tax.
Since the imperfectly competitive �rms enjoy rents, each government has a strategic
incentive to provide a competitive advantage to its domestic �rms so that they are
able to capture more foreign rent. Third, there is a �pollution shifting e¤ect� (or a
not-in-my-backyard e¤ect) that tends to increase tm, as each country tries to drive
polluting production from itself to the other countries.
The transboundary externality e¤ect and the pollution shifting e¤ect depend on
the extent to which pollution crosses jurisdictions. As increases from 0 to 13, the
former e¤ect becomes stronger while the latter e¤ect becomes weaker. Note that when
the good is clean (i.e. � = 0), both these e¤ects are non-existent. Moreover, the rent
capture e¤ect disappears when the market becomes competitive (i.e. as n!1). Theequilibrium pollution tax (10) in such a case becomes
limn!1
tm(�) j�=0 =
1
3�. (11)
As long as there is positive tari¤ (i.e. � = � > 0), each country enjoys tari¤ revenue on
imports and has to pay for exports. This gives them an incentive to substitute foreign
production for domestic production, and consequently to tax domestic �rms (the �tari¤
e¤ect�on the equilibrium tax). It is only when there is free trade (i.e. � = � = 0) as
well, that the equilibrium pollution tax rate in each country, (11), becomes zero.
In the symmetric equilibrium, substituting ti = tm(�) in (7) and (8), we have total
9
output produced equal to total output consumed in each country, so that its net import
is zero.9 The total output, produced or consumed, in each country is
Qi (�) =3a (1 + 2n)� 2� (1 + 3n)
3 ((1 + 2n) + � (1� 2 ) + 2n� (1� 3 ))
Multilateral tari¤ reduction increases output in each country as @Qi@�
< 0. Moreover,
the prohibitive tari¤ rate, at which imports become zero, is given by10
�mmax �3a (2n+ 1)
3�n (1� 2 + 2n (1� 3 )) + 6n2 + 9n+ 2
The impact of multilateral trade liberalization on the above-mentioned e¤ects, and
their trade-o¤, determines how the equilibrium pollution tax changes as a result. From
(10), we have
@tm(�)
@�=2
9n
1 + 3n+ 3n2 � 3n� (1� 2 + 2n (1� 3 ))2n+ 1 + � (1� 2 ) + 2n� (1� 3 ) (12)
From (12), Proposition 1 follows.
Proposition 1: Multilateral tari¤ reduction leads to an increase in the equilibriumpollution tax in each country (i.e. @t
mi
@�� 0) if and only if the pollution is su¢ ciently
harmful (i.e. � � 1+3n+3n2
3n(1�2 +2n(1�3 )) � �mA ).
Note that the threshold value of the pollution damage parameter, �mA ; itself depends
on the extent to which pollution is transboundary. As output increases and price falls
with trade liberalization, this increases pollution damage and decreases rents. Conse-
quently, a country�s incentive to raise tax to drive out polluting production increases,
and its incentive to lower tax to capture additional rents decreases. These exert an
upward pressure on the equilibrium pollution tax. However, a lower tari¤ also reduces
tari¤ revenues from imports. This reduces the country�s incentive to substitute foreign
production for domestic production by increasing the tax. As a result, equilibrium
pollution tax tends to decrease. The net impact on the tax depends on the relative
strength of the two counteracting forces.
Substitution of the equilibrium tax ti = tm(�) into (9) gives the equilibrium welfare
in each country,Wi = Wm(�). The e¤ect of multilateral trade liberalization on welfare
9This is similar to reciprocal dumping (Brander and Krugman, 1983).10That is, x2 = x3 = y1 = y3 = z1 = z2 � 0 if and only if � � �mmax.
10
is then given by
@Wm(�)
@�=
4 (3a� � ��� �) (3n+ 1)2
9 (� (2n (1� 3 ) + 1� 2 ) + 2n+ 1)2(13)
From (13), Proposition 2 follows.
Proposition 2: In the presence of transboundary pollution, multilateral tari¤ reductionleads to an increase in the welfare of each country (i.e. @W
m(�)
@�� 0) if and only if the
initial tari¤ rate � is su¢ ciently large (speci�cally � � 3a� 1+�
� �mA ).
The above threshold tari¤�mA is less than the prohibitive tari¤�mmax whenever n� <
13.
Whether tari¤ reduction, by increasing output, increases welfare of a country or
not depends on whether initially price exceeds marginal social cost of output in that
country. When pollution is cross-boundary, Proposition 2 indicates that welfare is non-
monotonic and concave in �. It is only when tari¤ is su¢ ciently high (� > �mA ), and the
associated output su¢ ciently low, that price exceeds marginal social cost. An increase
in output due to multilateral tari¤ reduction then increases welfare. The opposite
result holds when � < �mA . Thus, in the neighbourhood of free trade (when � ! 0),
multilateral tari¤ reduction decreases welfare in each country as long as � > 0.11
An implication of the non-monotonicity of Wm(�) is that the direction of change
in welfare of a country due to marginal multilateral reductions in tari¤may not be the
same as that due to a discrete jump in tari¤ (to free trade, for example). Note that
Wm(�) is non-monotonic and concave in � even when markets are competitive and
rents tend to zero.12 Existence of the rent capture e¤ect is, therefore, not necessary
for Proposition 2 to hold. In contrast, if the good was clean and the transboundary
externality and pollution shifting e¤ects were absent, multilateral trade liberalization
would always increase welfare.13 Furthermore, the threshold value of tari¤, �mA , is an
increasing function of the transboundary pollution parameter . In fact when pollution
is purely local (as in Burguet and Sempere, 2003), trade liberalization always improves
welfare as (13) implies @Wm(�)
@�j =0 = � 4�(3n+1)2
9(�+1)(2n+1)2< 0.
11We have, lim�!0
�@W
m(�)
@�
�= 4a� (3n+1)2
3(�(2n(1�3 )+1�2 )+2n+1)2 > 0
12As limn!1
�@W
m(�)
@�
�= 3a� �����
(1+��3� )2 :
13Again, @Wm(�)
@� j�=0 = � 4�(3n+1)2
9(2n+1)2< 0
11
4 Regional trade liberalization
We now move from multilateral trade liberalization, where all countries trading with
each other reduce their tari¤ rate, to regional trade liberalization where only a subset
of the trading countries reduce their tari¤ rate. Speci�cally, we analyze the impact of
an equal bilateral reduction in the tari¤ rate � by countries 1 and 2, with � remaining
unchanged. This could, for instance, happen if countries 1 and 2 sign a trade agree-
ment and decide to gradually move towards a regional free trade area, which does not
include country 3. Now we have an asymmetric case where the impact of regional trade
liberalization in the participating countries (countries 1 and 2) will di¤er from that in
the non-participating country (country 3).
The Nash equilibrium pollution taxes in each country, t1 (�;�) = t2 (�;�) and
t3 (�;�), are still subject to similar e¤ects as mentioned in Section 3. Although quali-
tatively similar, the magnitudes of these e¤ects di¤er across the countries when � 6= �.For instance, when � = 0 and the transboundary externality and pollution shifting
e¤ects are absent, the equilibrium pollution taxes are given by
ti (�;�) j�=0 =�(1� 6n2) + �B � A27n (3n+ 1) (2n+ 1)
; i = 1; 2
t3 (�;�) j�=0 =2�B � 2� (18n+ 42n2 + 27n3 + 2)� A
27n (3n+ 1) (2n+ 1);
whereA � 9a (2n+ 1) (3n+ 1) andB � 36n+78n2+54n3+5. Moreover, in the limitingcase when the number of �rms n!1 and the rent capture e¤ect is eliminated as well,
the equilibrium pollution taxes due to the tari¤ e¤ect are given by
limn!1
ti (�;�) j�=0 =1
3�; i = 1; 2 (14)
limn!1
t3 (�;�) j�=0 =1
3(2�� �) (15)
Lemma 1: A bilateral reduction in � tends to reduce equilibrium pollution tax in the
participating countries, and increase that in the non-participating country, through the
tari¤ e¤ect.
The above lemma follows from (14) and (15). Note that the tari¤ e¤ect on the
equilibrium pollution tax in the participating countries is independent of �, while that
in the non-participating country depends on both � and �. In particular, the tari¤
12
e¤ect encourages the non-participating country to subsidize production when its own
tari¤ � is (close to) zero, as limn!1�t3 (�;�) j�=0;�=0
�= �1
3� < 0.14
Substituting the Nash equilibrium taxes, t1 (�;�) = t2 (�;�) and t3 (�;�), into (7)
and (8) gives the equilibrium quantities xi (�;�), yi (�;�) and zi (�;�). Net imports
(i.e. consumption minus production) of country 1 and country 3 are respectively given
by
Q1 �X = n (x1 + y1 + z1)� n (x1 + x2 + x3) =1
3
(�� �)�(3n+ 1)2 � 3n�C
�(3n+ 1) (1 + 3�C)
Q3 � Z = n (x3 + y3 + z3)� n (z1 + z2 + z3) =2
3
(�� �)�(3n+ 1)2 � 3n�C
�(3n+ 1) (1 + 3�C)
where C � (1� 3 ) (2n (1� 3 ) + 1� 2 ) � 0. Net imports are negative (positive) inthe participating countries (non-participating country) when � < � and � is su¢ ciently
small. Moreover, we have
@X
@�=@Y
@�< 0;
@Q1@�
=@Q2@�
< 0;@Z
@�> 0;
@ (X + Y + Z)
@�< 0 (16)
@Q3@�
=2
9
3n� (1� 2 ) + 6n2� (1� 3 )� 1� 3n� 3n2n (3n+ 1) (1 + 2n+ � (2n (1� 3 ) + 1� 2 )) (17)
From (16) and (17), Lemma 2 follows.
Lemma 2: Regional trade liberalization increases production and consumption in theparticipating countries but decreases production in the non-participating country. More-
over, consumption in the non-participating country increases if and only if the pollution
damage parameter is su¢ ciently small, i.e. � < 1+3n+3n2
3n(1�2 )+6n2(1�3 ) :
The changes in consumption (or sales) in the participating and non-participating
countries lead to corresponding changes in the price of the polluting good, and the
strength of the rent capture e¤ect, in these countries. The increase in sales in the
participating countries implies a decrease in the price of the polluting good in these
countries. This weakens the rent capture e¤ect in the non-participating country and
tends to raise its pollution tax. If � is su¢ ciently small, regional trade liberalization
also increases sales and decreases price in the non-participating country. This tends
14Recall that, given our assumption of constant emission intensity of output, the tax on pollutionis equivalent to a production tax.
13
to raise pollution tax in the participating countries through the rent capture e¤ect.
On the other hand, if � is large enough to decrease sales and increase price in the
non-participating country, the rent capture e¤ect in each participating country is then
subject to counteracting forces. For example, country 1�s rent capture e¤ect tends to
become weaker due to the fall in price in country 2 but stronger due to the rise in price
in country 3.
In general, the above-speci�ed changes in consumption and production drive the
changes in equilibrium pollution tax and welfare that arise due to regional trade lib-
eralization. For expositional ease, we examine these changes for the two polar cases
where pollution is purely local and perfectly transboundary.
4.1 Local pollution
This subsection examines the case of local pollution when = 0. The impact of an
equal bilateral reduction in tari¤ by countries 1 and 2 on equilibrium pollution taxes
is then given by
@ti (�;�)
@�j =0 =
M
27n (� + 1) (2n+ 1) (3n+ 1) (3� + 6n� + 1); i = 1; 2 (18)
whereM � 3 (2n+ � + 2n�) (13n+ 9n2 + 3)�27n�2 (3n+ 1) (2n+ 1)2+18n+5, and
@t3 (�;�)
@�j =0 =
2
27
3n� (2n+ 1) (18n+ 27n2 + 2)� (18n+ 42n2 + 27n3 + 2)n (� + 1) (2n+ 1) (3n+ 1) (3� + 6n� + 1)
(19)
Proposition 3 follows from the above.
Proposition 3: When pollution is purely local, regional trade liberalization increasesequilibrium pollution tax in the participating (non-participating) country if and only if
the pollution damage parameter � is su¢ ciently large (small).
proof: From (18) we have that M is concave in �; M j�=0 > 0, and M � 0 if and
only if � � �rA � 13n+9n2+3+p138n+835n2+2466n3+3537n4+1944n5+9
18n(2n+1)(3n+1). Moreover, from (19),
@t3(�;�)@�
j =0 � 0 if and only if � � �rB � 18n+42n2+27n3+23n(2n+1)(18n+27n2+2)
.
Proposition 3 indicates that, for pollutants that have very harmful but localized
impacts, regional trade liberalization is likely to strengthen environmental protection
in participating countries but weaken such protection in non-participating countries.
14
Regional trade liberalization increases polluting production in the participating
countries. Consequently, in these countries, the changing strength of the pollution
shifting e¤ect tends to increase the equilibrium pollution tax, whereas the change in
the tari¤ e¤ect tends to decrease it. When pollution is purely local and the pollution
shifting e¤ect is at its strongest, Proposition 3 shows that the equilibrium pollution
tax in the tari¤-reducing countries increases for su¢ ciently high values of �. On the
other hand, in the non-participating country, the tari¤ e¤ect tends to increase its
pollution tax. Moreover, since regional tari¤reduction decreases production in the non-
participating country, the pollution shifting e¤ect operates in the opposite direction in
this country. Hence, pollution tax increases in the non-participating country when �
is su¢ ciently small.
The equilibrium welfare of each country, denoted by Wi (�;�), can be derived by
substituting the equilibrium pollution taxes ti (�;�) into (9). The impact of regional
trade liberalization on each country�s welfare is derived as follows. We �rst di¤erentiate
Wi (�;�) with respect to �, and then evaluate the resulting expression, which is a
function of both � and �, at � = � = 0. This gives the impact of marginal bilateral
tari¤ reduction by countries 1 and 2 on welfare of country i, when evaluated in the
neighbourhood of free trade, as follows:
@Wi (�;�)
@�j =0; �=�=0 = �
an
3 (3n+ 1) (� + 1)< 0, i = 1; 2 (20)
@W3 (�;�)
@�j =0; �=�=0 =
2an
3 (3n+ 1) (� + 1)> 0 (21)
From (20) and (21), we have Proposition 4.
Proposition 4: When pollution is purely local, regional trade liberalization in theneighbourhood of free trade increases (decreases) equilibrium welfare in the participating
(non-participating) country.
The changes in welfare in the participating and non-participating countries follow
from the changes in production and consumption in these countries brought about by
regional trade liberalization. For instance, with local pollution, price of the polluting
good exceeds its marginal social cost in the participating countries. Consequently, the
increase in consumption and production in these countries increases their welfare.
15
4.2 Perfectly transboundary pollution
We now take up the case of perfectly transboundary pollution where pollution gen-
erated in each country a¤ects all three countries equally (i.e. = 13). As a result,
the pollution shifting e¤ect is absent �this approximates the situation with respect to
global pollutants such as greenhouse gases. The impact of an equal bilateral reduction
in tari¤ rate by countries 1 and 2 on equilibrium pollution taxes is now given by
@ti (�;�)
@�j = 1
3=36n+ � + 4n� + 78n2 + 54n3 + 5
9n (3n+ 1) (6n+ � + 3)> 0; i = 1; 2 (22)
@t3 (�;�)
@�j = 1
3= �2 (18n+ � + 7n� + 42n
2 + 27n3 + 9n2� + 2)
9n (3n+ 1) (6n+ � + 3)< 0 (23)
From (22) and (23), Proposition 5 follows.
Proposition 5: When pollution is perfectly transboundary, regional trade liberalizationdecreases (increases) equilibrium pollution tax in the participating (non-participating)
country.
When pollution is perfectly transboundary, the increased production in the par-
ticipating countries does not exert an upward pressure on their pollution tax through
the pollution shifting e¤ect. Instead, the reduction in � decreases pollution tax in the
participating countries through the tari¤ and rent capture e¤ects. Converse changes
in the tari¤ and rent capture e¤ects move the pollution tax rate upwards in the non-
participating country, as � decreases.
The equilibrium welfare level in each country is derived by substituting the equilib-
rium pollution taxes ti (�;�) into (9). The impact of regional trade liberalization on
each country�s welfare is obtained as earlier. Thus, we �rst di¤erentiate Wi (�;�) with
respect to �, and then evaluate the resulting expression, which is a function of both �
and �, at � = � = 0. This yields:
@Wi (�;�)
@�j = 1
3; �=�=0 =
a (� (6n+ 33n2 + 54n3 + 1)�G)3 (3n+ 1) (6n+ � + 3)2
; i = 1; 2 (24)
@W3 (�;�)
@�j = 1
3; �=�=0 =
2a (� (48n+ 129n2 + 108n3 + 5) +G)
3 (3n+ 1) (6n+ � + 3)2> 0 (25)
where G � 9n (2n+ 1)2 + �2 (7n+ 9n2 + 1). The following result holds.
16
Proposition 6: When pollution is perfectly transboundary, regional trade liberalizationin the neighbourhood of free trade (i) decreases equilibrium welfare in the participating
country if and only if the pollution damage parameter takes intermediate values (i.e.
� 2��L; �H
�); and (ii) decreases equilibrium welfare in the non-participating country.
proof: Follows from (24) and (25). De�ne the numerator of (24),N � � (6n+ 33n2 + 54n3 + 1)�G. We have N is concave in �, N j�=0 < 0, and N � 0 if and only if 0 < �L � � � �H
where��L; �H
�� 6n+33n2+54n3+1�
p(108n3�24n2�33n+1)(3n+1)3
2(7n+9n2+1):
The above changes in welfare follow from the relevant changes in production and
consumption brought about by regional trade liberalization in the participating and
non-participating countries.
5 Comparison
Table 1 summarizes the results for multilateral and regional trade liberalization in
the neighbourhood of free trade. The results emerge from the interaction of the rent
capture, pollution shifting and tari¤ e¤ects, and changes in these e¤ects depend on
the nature of the trade liberalization process. While multilateral trade liberalization
increases production and consumption in all countries, regional trade liberalization re-
duces production in the non-participating country and may decrease its consumption
as well (the latter happens when the pollution damage parameter is su¢ ciently large).
The changes in consumption and production have corresponding impacts on the rent
capture and pollution shifting e¤ects in these countries. Moreover, the tari¤ e¤ect
reduces equilibrium pollution tax in all countries in the multilateral case, and in par-
ticipating countries in the regional case, but increases the non-participating country�s
tax in the case of regional trade liberalization.
A comparison of the results in Table 1 reveals that multilateral and regional trade
liberalization a¤ect pollution tax and welfare in the tari¤-reducing countries in sim-
ilar ways when pollution is local, but in dissimilar ways for perfectly transboundary
pollution.
Multilateral trade liberalization increases pollution tax in the participating coun-
tries if and only if pollution is su¢ ciently harmful. A similar result holds for regional
trade liberalization when pollution damage is localized. However, for pollutants that
have large inter-jurisdictional spillovers (e.g. greenhouse gases), regional trade liberal-
17
ization always decreases the participating countries�pollution tax. When pollution is
transboundary, regional trade liberalization is thus more likely than multilateral trade
liberalization to weaken environmental protection in the tari¤-reducing countries.
Moreover, when pollution is transboundary, countries undertaking multilateral trade
liberalization in the neighbourhood of free trade experience a reduction in welfare. In
contrast, regional trade liberalization, in the neighbourhood of free trade, may increase
welfare in the participating countries. The di¤erence between multilateral and regional
trade liberalization in terms of their impact on welfare of the tari¤-reducing countries
suggests that multilateralism is likely to be the less preferred option for the establish-
ment of free trade amongst countries. This is in line with the recent experience with
respect to multilateral and regional trade agreements discussed in the Introduction.
Further, regional trade liberalization is found to a¤ect pollution tax and welfare
in the participating and non-participating countries in very dissimilar ways. When
pollution is local but su¢ ciently harmful, regional trade liberalization strengthens en-
vironmental protection in participating countries but weakens such protection in the
non-participating country.
6 Conclusion
The paper considers trade between identical countries with imperfectly competitive
markets and examines the impact of multilateral and regional tari¤ reduction on their
non-cooperative equilibrium pollution tax and welfare. We �nd that when production
causes pollution, and countries impose a strategic pollution tax that adjusts to changing
rates of tari¤, trade liberalization may weaken environmental policy and reduce welfare
in the tari¤-reducing countries.
When a pollution tax is chosen strategically by countries, it is a¤ected by the rent
capture, pollution shifting and tari¤ e¤ects. Tari¤ reduction changes the magnitude of
these e¤ects and is observed to a¤ect equilibrium pollution tax and welfare in ways that
depend, in part, on the characteristics of the pollution itself (viz. pollution damage
parameter and pollution spillover parameter).
The nature of the trade liberalization process matters as well. While the exist-
ing literature has focused on analyzing multilateral trade liberalization, much tari¤
reduction in the world is undertaken through regional trade agreements amongst a
sub-group of trading countries. We �nd that multilateral and regional trade liberaliza-
18
tion a¤ects pollution tax and welfare in the tari¤-reducing countries in similar ways for
local pollution but in dissimilar ways for perfectly transboundary pollution. As well,
when pollution is transboundary, regional trade liberalization is more likely than mul-
tilateral trade liberalization to weaken environmental protection in the tari¤-reducing
countries.
In their analysis of trade with perfect competition and non-strategic environmental
policy, Copeland and Taylor (2004, p. 29) had noted that, �If pollution externalities are
fully internalized, trade must always increase welfare.�However, in our model involving
a �exible environmental tax that not only adjusts to changing openness to trade but is
also strategically distorted, we �nd that the distortions can be severe enough to negate
the bene�cial e¤ects of freer trade. Our welfare results thus provide formal support
to what Kennedy (1994, p. 62) had speculated, �However, one can speculate that
the strategic distortions associated with openness have the potential to be su¢ ciently
destructive as to more than o¤set any bene�ts associated with trade liberalization.�
We have examined the analytically tractable case of identical countries facing lin-
ear demand and quadratic pollution cost. Nevertheless, even when these simplifying
assumptions are relaxed, production, consumption, and the various strategic e¤ects de-
scribed above will continue to change di¤erently across countries, leading to di¤erential
impacts of multilateral and regional trade liberalization on environmental protection
and social welfare.�
19
References
Antweiler, W., B. Copeland, and M.S. Taylor (2001), �Is Free Trade Good for the
Environment�, American Economic Review, 91, 877-908.
Baksi, S. and A. Ray Chaudhuri (2009), �On Trade Liberalization and Transbound-
ary Pollution�, Economics Bulletin, 29(4), 2599-2606.
Barrett, S. (1994), �Strategic Environmental Policy and International Trade�, Jour-
nal of Public Economics, 54, 325-338.
Benarroch, M. and H. Thille (2001), �Transboundary Pollution and the Gains from
Trade�, Journal of International Economics, 55, 139-159.
Brander, J. and P. Krugman (1983), �A �Reciprocal Dumping�Model of Interna-
tional Trade,�Journal of International Economics, 15, 313-321.
Burguet, R. and J. Sempere (2003), �Trade Liberalization, Environmental Policy,
and Welfare�, Journal of Environmental Economics and Management, 46, 25-37.
Copeland, B. and M.S. Taylor (1999), �Trade, Spatial Separation, and the Envi-
ronment�, Journal of International Economics, 47, 137-168.
Copeland, B. and M.S. Taylor (2003), Trade and the Environment: Theory and
Evidence, Princeton: Princeton University Press.
Copeland, B. and M.S. Taylor (2004), �Trade, Growth, and the Environment�,
Journal of Economic Literature, 42(1), 7-71.
Ederington, J. and J. Minier (2003), �Is Environmental Policy a Secondary Trade
Barrier? An Empirical Analysis�, Canadian Journal of Economics, 36(1), 137-154.
Greenstone, M. (2002), �The Impacts of Environmental Regulations on Industrial
Activity: Evidence from the 1970 and 1977 Clean Air Act Amendments and the Census
of Manufactures,�Journal of Political Economy, 110(6), 1175-1219.
Kennedy, P. (1994), �Equilibrium Pollution Taxes in Open Economies with Imper-
fect Competition�, Journal of Environmental Economics and Management, 27, 49-63.
Krutilla, K. (1991), �Environmental Regulation in an Open Economy�, Journal of
Environmental Economics and Management, 20, 127-142.
Managi, S., A. Hibiki, T. Tsurumi (2009), �Does Trade Openness Improve Environ-
mental Quality?�Journal of Environmental Economics and Management, 58, 346-363.
McAusland, C. (2008), �Trade, Politics, and the Environment: Tailpipe vs. Smoke-
stack,�Journal of Environmental Economics and Management, 55, 52-71.
Ulph, A. (1996), �Environmental Policy and International Trade when Govern-
ments and Producers Act Strategically,� Journal of Environmental Economics and
20
Management, 30, 265-281.
OECD (2007), �Environment and Regional Trade Agreements,�ISBN: 9789264006652.
21
22
Figure 1: Production, Consumption, and Tariff rates for countries 1, 2 and 3
1 (X, Q1)
φ
Φ Φ
2 (Y, Q2)
3 (Z, Q3)
23
Table 1: Effects of multilateral vs. regional tariff reduction on pollution tax and welfare (in the neighbourhood of free trade)
Multilateral Trade Liberalization Regional Trade Liberalization
0γ = 13(0, ]γ ∈ 0γ = 1
3γ =
1t iff is largeβ↑ iff is largeβ↑ iff is largeβ↑ ↓
3t iff is largeβ↑ iff is largeβ↑ iff is largeβ↓ ↑
1W ↑ ↓ ↑ iff is intermediateβ↓
3W ↑ ↓ ↓ ↓